Saw many examples on the net, but they are all using tensorflow2.0 training images, the results are also tells readers the model through several epoch, result how, but no one introduction, training good model after the how to apply this model, also is the place that I have been very confused, some people should be to want to take advantage of tensorflow realization of image recognition, first founded model training data set, after the training to forecast the new photos again, but it is not to find the corresponding data, also has the corresponding class MOOC, one day a little mention, but still the old version, a lot of usage are different, such as Sess, so now don't know how to continue, and model training is good, but I do not know how to use